And Then You're Dead

The Autopsy / Case File · Medical Error & Preventable Care

Four Drug Classes

Ask a hospital pharmacist which medications land the most older adults in the emergency room, and the honest answer disappoints anyone expecting an exotic culprit. It isn't a rare interaction or an inappropriate prescription pulled from a watch list. It's blood thinners and diabetes medication — drugs tens of millions of older Americans take exactly as directed, every single day.

A national surveillance study published in the New England Journal of Medicine found that an estimated 99,628 emergency hospitalizations occur each year among U.S. adults 65 and older because of adverse drug events — and that four medications or medication classes, alone or in combination, accounted for 67 percent of them: warfarin (33.3 percent), insulins (13.9 percent), oral antiplatelet agents (13.3 percent), and oral hypoglycemic agents (10.7 percent). Nearly two-thirds of those hospitalizations, 65.7 percent, resulted from unintentional overdose — not a prescribing error, in most cases, but the ordinary, expected risk of a drug doing exactly what it's designed to do, taken by a body with less margin for error than it used to have.

And Then You're Dead · September 2026 · 10 min read

Join the discussion →
A weekly plastic pill organizer with compartments labeled Sunday through Saturday, filled with various pills and capsules, next to loose tablets and capsules on a counter
A weekly pill organizer, the everyday tool millions of older Americans use to manage exactly the kind of medications — anticoagulants, insulin, antiplatelet agents, oral diabetes drugs — that a landmark national study found behind two-thirds of emergency hospitalizations for drug harm in adults 65 and older. Photo: Stevepb, CC0, via Wikimedia Commons.

The instinct, when picturing a dangerous medication error in an older patient, tends toward something dramatic: a look-alike, sound-alike drug mix-up; a dose miscalculated by a rushed nurse; a dangerous interaction between two drugs that never should have been combined. Those things happen, and hospitals track them as distinct patient-safety categories. But the largest single driver of emergency hospitalizations for drug harm in older Americans, according to the most rigorous national surveillance data available, is none of them. It's the correctly prescribed, correctly filled, correctly taken medication — specifically, one of a short list of drugs whose entire therapeutic purpose requires operating close to a line between "working" and "dangerous," in a population whose kidneys, liver, and overall physiological reserve have less capacity to keep them on the right side of it than a younger patient's would.

The Study Behind the Number

The benchmark data comes from Daniel S. Budnitz and colleagues at the CDC, published in the New England Journal of Medicine on November 24, 2011, under the title "Emergency Hospitalizations for Adverse Drug Events in Older Americans." The study drew on a nationally representative surveillance system — the CDC's National Electronic Injury Surveillance System-Cooperative Adverse Drug Event Surveillance Project, which captures detailed case data from a representative sample of U.S. emergency departments — covering the years 2007 through 2009. Extrapolated nationally, the surveillance data produced an estimate of 99,628 emergency hospitalizations per year among adults 65 and older specifically attributable to adverse drug events, out of a much larger total of ED visits for drug harm in that age group (many of which didn't require hospitalization). Of the hospitalizations, 65.7 percent were coded as resulting from unintentional overdose — meaning the harm arose from the drug's known, expected effect occurring at a dose or in a physiological context where that effect became dangerous, not typically from a dispensing or administration error.

The four-drug finding is what made the study a lasting benchmark rather than just another prevalence estimate. Warfarin alone accounted for 33.3 percent of hospitalizations — the single largest share by a wide margin — followed by insulins at 13.9 percent, oral antiplatelet agents (drugs like clopidogrel, taken to prevent blood clots) at 13.3 percent, and oral hypoglycemic agents (the pill-form diabetes medications distinct from insulin) at 10.7 percent. Combined, alone or in combination with each other, those four classes were implicated in 67.0 percent of all the hospitalizations the study identified. The researchers' explicit conclusion — that most emergency hospitalizations for adverse drug events in older adults trace back to a small number of commonly used medications, not the drugs typically flagged as high-risk or inappropriate for older patients on standard prescribing checklists like the Beers Criteria — reframed how health systems were supposed to think about medication safety in this population. The danger wasn't concentrated in the drugs doctors were already trained to double-check. It was concentrated in the drugs so common that nobody thought to.

0
percent of ER hospitalizations for adverse drug events in adults 65+ caused by just four drug classes
0
estimated annual U.S. emergency hospitalizations for adverse drug events among adults 65 and older

Why These Four, Specifically

The pharmacology behind the finding isn't obscure — it's the reason these drugs are prescribed at all. Warfarin works by thinning blood enough to prevent dangerous clots, which means its effective dose and its bleeding-risk dose sit close together, and that margin narrows further with age as kidney and liver function decline, interacting medications accumulate, and diet and hydration status fluctuate more than they did decades earlier. Insulin and oral hypoglycemic agents work by lowering blood sugar, and the line between "controlled" and "dangerously low" — hypoglycemia severe enough to cause confusion, seizure, or loss of consciousness — is a matter of degree, not kind, especially in patients whose kidney function affects how long a dose of insulin or a sulfonylurea stays active in the bloodstream. Oral antiplatelet agents share warfarin's basic bleeding-risk mechanism through a different pathway. All four drug classes, in other words, are dangerous for the same structural reason they're effective: they're titrated to operate near a therapeutic edge, in a population where that edge shifts unpredictably as organ function changes, sometimes month to month, in ways a fixed prescription doesn't automatically track.

That's a meaningfully different problem than the one polypharmacy discussions often default to — the idea that older patients are harmed mainly by taking too many drugs at once, with dangerous combinations multiplying as prescription counts climb. Drug interactions and cumulative polypharmacy burden are real, documented hazards in their own right. But the Budnitz findings point to something narrower and, in some ways, harder to solve through simplification alone: even a patient on exactly the medications their conditions require, at the correct doses, prescribed by a careful physician, remains at meaningful risk from these four classes specifically, because the risk is inherent to what the drugs do, not a symptom of prescribing too many of them. A patient on warfarin and nothing else can still bleed dangerously. Cutting their total pill count wouldn't change that.

These aren't the drugs on a "high-risk medication" list handed out at hospital discharge. They're the drugs already in the weekly pill organizer on the kitchen counter, taken exactly as prescribed, by design operating close enough to the edge that "exactly as prescribed" is sometimes still not enough.

What's Changed, and What Hasn't, Since 2011

The Budnitz findings have held up as a benchmark in part because the underlying pharmacology hasn't changed and in part because follow-up surveillance has kept confirming the same basic pattern rather than overturning it. The same CDC research group's later work, including a 2021 JAMA study on medication-related emergency department visits, has continued to find anticoagulants and diabetes medications disproportionately represented among the drugs driving the most severe outcomes in older patients, even as the specific drugs within those classes have shifted — newer direct oral anticoagulants have partially displaced warfarin in prescribing patterns since 2011, marketed in part on a somewhat more predictable bleeding-risk profile and less need for the frequent blood monitoring warfarin requires. Whether that shift has meaningfully reduced the anticoagulant-related hospitalization share nationally is still an active question in the more recent medication-safety literature, complicated by the fact that newer anticoagulants carry their own distinct bleeding risks and, unlike warfarin, don't have a routine blood test that flags a patient drifting toward danger before a bleeding event actually happens.

What the original study's core policy implication changed, more than the drugs themselves, was where health systems have since tried to intervene. Rather than treating "reduce the pill count" as the primary safety strategy, medication-safety programs increasingly focused on the specific mechanics of these four drug classes: standardized warfarin monitoring protocols and dosing algorithms, structured insulin-dosing safeguards in hospitals and long-term care settings, and discharge processes specifically built around anticoagulant and diabetes-medication education, since patients transitioning out of a hospital stay are at elevated risk of dosing confusion during exactly the period when their regimen is most likely to have just changed. The intervention target shifted from "how many drugs" to "these four, specifically, monitored more carefully" — a narrower, more targeted response than a blanket polypharmacy-reduction push, and one directly traceable to what the 2011 data actually showed rather than to the more general anxiety about older patients' pill counts that preceded it.

The Discharge Window

A specific detail in the broader medication-safety literature makes the case for targeted intervention over blanket polypharmacy reduction even sharper: adverse drug events in older adults cluster disproportionately in the weeks immediately following a hospital discharge, when a patient's regimen has often just changed — a new anticoagulant started, an insulin dose adjusted for a condition that prompted the hospitalization in the first place, a medication list reconciled across multiple prescribers who may not have been fully coordinated during the stay. That transition window is precisely when the four highest-risk drug classes are most likely to be new, recently adjusted, or unfamiliar to a patient and caregiver managing the change at home without the same oversight a hospital floor provides. It's also, not coincidentally, the exact window discharge-education programs focused on anticoagulant and diabetes-medication safety were built to target once the 2011 findings reframed where the risk actually concentrated.

And Then You're Dead

Nobody pictures a blood thinner or an insulin pen as the dangerous drug in the house. The opioid bottle gets the warning label; the antipsychotic gets the black-box alert. Warfarin and insulin just sit in the weekly organizer, doing exactly what they're supposed to do, in a body with less room than it used to have for exactly right to still be safe.

That's it. That's the whole thing.

Comments

Loading…

Comments are anonymous and moderated.

Sources